Myosoft: An automated muscle histology analysis tool using machine learning algorithm utilizing FIJI/ImageJ software

Area of Science:

• Muscle physiology and histology analysis
• Biomedical image analysis software development

Background:

• Accurate quantification of myofiber size and type is crucial for understanding muscle function and disease.
• Manual analysis of muscle histology is time-consuming and labor-intensive.

Purpose of the Study:

• To introduce Myosoft, a novel software tool for high-throughput muscle histology analysis.
• To demonstrate the accuracy, speed, and robustness of Myosoft in analyzing myofiber characteristics.

Main Methods:

• Muscle sections were stained for cell boundaries (laminin) and myosin heavy chain isoforms.
• Myosoft software, integrated with FIJI (ImageJ), was employed for automated analysis of gastrocnemius/soleus muscle transverse sections.
• Analysis included myofiber size and type determination.

Main Results:

• Myosoft achieved accurate analysis of hundreds to thousands of muscle fibers within 25 minutes.
• The software demonstrated a speed advantage of over 10 times compared to manual analysis.
• Myosoft effectively handled high-content images with suboptimal quality, outperforming comparable existing programs.

Conclusions:

• Myosoft is a reliable, accurate, and convenient tool for high-content muscle histology analysis.
• The software offers a significant improvement in throughput and efficiency for researchers.
• Myosoft is freely available for download from Github.